Capacitance Immunosensors for the early detection
of Circulating Cancer Biomarkers
Contents
Abstract - English
8
Abstract - Italiano
10
1 Introduction
12
2 Materials and Methods
20
2.1 Device Fabrication &
Mea-suring Setup . . . .
20
2.2 Oligonucleotides . . . . .
27
2.3 An insight on SAM
for-mation
. . . .
29
2.4 ssDNA-SAM Density .
.
33
2.5 Experimental procedures
36
2.5.1
De-hybridization
pro-tocols for device
re-generation . . . .
37
3 A Biosensor for Direct
Detection of DNA Sequences
38
3.1 DNA-SAM Electrical
Char-acterization . . . .
38
3.1.1
The eects of the
Applied Potential
on the ssDNA-SAMs 39
3.1.2
Divalent salt eects
CONTENTS 4
3.2 The biophysics of DNA
hybridization .
. . . .
44
3.3 Kinetics and Dynamics of
DNA Hybridization .
. .
49
3.4 DNA Hybridization as a
Function of the Applied
Parameters . . . .
53
3.5 Mismatch detection in DNA
monolayer . . . .
60
4 DNA Surface Hybridization
via Theoretical Model
65
4.1 The Capacitive Model . .
66
4.2 The Electrokinetics
com-ponent of the Model . . .
70
4.2.2
DEP force
calcu-lation . . . .
74
4.2.3
ACEO driving and
quantication
. .
74
4.3 The nal approach .
. . .
79
5 Detecting miRNAs relevant
for heart failure disease
82
5.1 Quantication of free
cir-culating miRNA in
cellu-lar extract
. . . .
83
5.2 Device specicity in
Hu-man Plasma . . . .
87
5.3 miRNA detection in
CONTENTS 6
5.3.1
The Argonaute
fac-tor
. . . .
89
5.3.2
Circulating miRNA
concentration in
Hu-man Plasma . . .
93
6 Protein Detection: HER2
(Breast Cancer Biomarker)
96
6.1 ECD-HER2 Detection .
.
98
6.1.1
Nanobody
conju-gation via Maleimide
reaction . . . .
99
8 Perspectives (3 Co-planar
electrodes setup, AuNPs)
106
8.1 Gold Nanoparticles
mod-ied gold working electrode 107
8.2 3 Co-planar electrodes setup
as a possible
implemen-tation .
. . . .
110
Appendices117 Appendix A117
Abstract - English
The quantication of signicant amounts of disease biomarkers circulating in the blood-stream represents one ofthe challenging frontiers in biomedicine. The complexity of blood composition has opened the quest for novel detection technologies, capable of dis-cerning small amount of specic biomarkers from other blood proteins/oligonucleotides and of reliably measuring them. In this context, we have developed a device based on dierential double-layer capacitance readout at microfabricated gold electrodes and demonstrated its detection performance in real bio-sample volumes.
In particular, in this PhD thesis, I will show my results in improving and implement-ing a biosensor based on a three-electrodes electrochemical readout usimplement-ing two miniatur-ized gold (working and counter) electrodes and a mm-sminiatur-ized AgClpellet reference elec-trode, with the nal scope of detecting cancer biomarkers circulating in blood in real-time and in-situ. This biosensor has demonstrated to be able to detect DNA-hybridization with a detection limit of 1 pM, starting from a probing ssDNA-SAM (self-assembled monolayer) on the gold-coated working electrode.The measurements were rst carried out in pure saline buer solution, monitoring the dierential capacitance at the Working Electrode versus the incubation time.The kinetics studies, modeled using the Langmuir adsorption isotherm, not only give us important information on DNA hybridization ki-netics but also allow to detect eventual mismatches along the target DNA sequence proving to be sensitive to the position of the mismatch with respect to the surface of the device. Furthermore, we are able to detect and quantify, in human extract and plasma, an unknown concentration of a specic microRNA (miRNA) biomarker connected to heart failure disease,using miRNA/complementary ssDNA calibration curves. The re-sults were then conrmed using a real time qPCR by our MD partners at the University of Udine, D. Cesselli and A.P. Beltrami.
Beside miRNA, I have demonstrated that my device can detect more complex com-ponents such as protein biomarkers on single-domain antibodies (e.g.VHH fragments)
DNA conjugates in human serum. In particular, I have focused on the detection of the protein HER2, overespressed in many types of cancer.In addition, I have performed a systematic characterization of the device varying the physiological conditions (e.g.salt type: KCl, NaCl, MgCl2, PBS, etc.and concentration),at dierent ssDNA SAM den-sity conditions (estimated by using an XPS) and applying dierent potential in order to have a more comprehensive understanding of the phenomena occurring at the elec-trode/electrolyte interface. Such studies have led to the implementation of a theoretical model, to provide an explanation, consistent with the experimental data, of the biophys-ical phenomena that contribute to the biorecognition of the events of interest.Recently, I started exploring new routes to improve the sensitivity of our devices: on one hand, through Au nanoparticles amplication; on the other hand, through the development of a multiplexing system, which also requires the miniaturization of the reference electrode on the same plane of the other two gold electrodes, to increase device portability and its limit of detection.
Abstract - Italiano
La detection e misurazione di biomarcatori tumorali e non circolanti nel sangue umano rappresenta una delle nuove e più impegnative frontiere in campo biomedico.Tuttavia, la complessità delsangue umano ha reso necessario lo sviluppo dinuove tecnologie di rilevamento, in grado di discernere piccole quantità dibiomarcatori specici dagli altri oligonucleotidi o componenti proteiche presenti nel sangue e di misurarle in modo ad-abile. Ed è in questo contesto, che abbiamo sviluppato un biosensore in grado di misurare le variazioni della capacità dierenziale che avvengo all'interfaccia elettrodo/elettrolita
(per questo nota come capacità di double-layer ) e connessi ad eventi di bioriconoscimento dimostrando buone capacità di rilevamento anche in ambienti reali quali plasma umano ed estratto cellulare.
In particolare, in questa tesi di dottorato, vi mostrerò i risultati da me ottenuti nel miglioramento di un dispositivo costituito da un redout elettrochimico a due elettrodi d'oro microfabbricati (di lavoro e ausiliario) e un elettrodo di riferimento di AgCl di dimensioni millimetriche, con la scopo nale di rilevare biomarcatori tumorali circolanti nel sangue, in tempo reale e in-situ.Questo biosensore ha dimostrato di essere in grado di rilevare il processo di ibridazione del DNA con un limite di detection di 1 pM, a partire da un monostrato di DNA a singolo lamento (ssDNA-SAM) auto-assemblato sull'elettrodo di lavoro. Le misurazioni sono state eettuate in una soluzione tampone salina pura e la capacità dierenziale in corrispondenza dell'elettrodo di lavoro in funzione del tempo di incubazione è stata poi monitorata.
Gli studi cinetica, attraverso il modello di adsorbimento di Langmuir, non solo ci danno informazioni importanti sulla cinetica di ibridazione DNA ma anche permettere di rilevare eventuali spaiamenti di basi lungo la sequenza del DNA d'interesse dimostrando, inoltre, di essere sensibile alla posizione del disallineamento rispetto alla supercie il dis-positivo. Inne, siamo in grado di rilevare e quanticare, sia in estratto cellulare che in plasma umano, una concentrazione sconosciuta di uno specico frammento di RNA
croRNA, o miRNA) biomarcatore specico dell'insucienza cardiaca, utilizzando curve di calibrazione di ssDNA/miRNA complementare. I risultati sono stati poi confermati attraverso l'uso di una qPCR in tempo reale grazie ai nostri partner presso l'Università di Udine, i MD D. Cesselli e A.P. Beltrami e collaboratori.
Oltre lo studio e rilevazione dei miRNA, ho dimostrato che il dispositivo è in grado di rilevare componentipiù complessiquali biomarcatori proteici su anticorpi a singolo dominio (ad esempio frammenti VHH) coniugati a lamenti di DNA in siero umano. In particolare, ci siamo concentrati sulla rilevazione della proteina HER2, la cui sovrae-spressione è stata rilevata in molti tipi di cancro ma in special modo in forme par-ticolarmente aggressive dicancro al seno. Inoltre, ho eseguito una caratterizzazione sistematica del dispositivo variando le condizioni siologiche (ad esempio il tipo di sale: KCl, NaCl, MgCl2, PBS, etc. e lo loro concentrazione),a diverse condizionidi densità del ssDNA-SAM (stimata utilizzando un XPS) e applicando diversi potenziali in modo da avere una comprensione più completa dei fenomeni che avvengono all'interfaccia elet-trodo/elettrolita. Tali studi hanno portato alla realizzazione di un modello teorico, in grado di fornire una spiegazione, coerente con i dati sperimentali, dei fenomeni biosici che contribuiscono al bioriconoscimento degli eventi di interesse.
Recentemente, ho iniziato a esplorare nuove vie per migliorare la sensibilità dei nostri dispositivi: da un lato, attraverso l'uso di nanoparticelle d'oro in grado di amplicare il segnale delle variazionid'interesse; dall'altro, attraverso lo sviluppo di un sistema di misurazioni a multicanale, che richiede anche la miniaturizzazione dell'elettrodo di riferimento sullo stesso piano degli altri due elettrodi in oro, al ne di aumentare la portabilità del dispositivo e inoltre il suo limite di rilevazione.
Chapter 1
Introduction
In the last decades the growing interest towards personalized therapies has motivated the development of an increasing number of miniaturized, label-free devices to be used as fast diagnostic tools for medical treatment ([1, 2, 3, 4, 5]).Moreover, in recent years the measurements of blood (and its components) impedance through an alternating current has been suggested as a non-invasive approach to determine some blood disorders.[6]. The nal goal, of all these devices, is the detection of biological biomarkers (as proteins, microRNA (miRNA), small pieces of DNA, etc.) circulating in the bloodstream, and to selectively distinguish them from the huge amount of other,non-disease-representative molecules.
Label-free biosensors require only a single recognition element, leading to simplied assay design,decreased assay time and reduction in reagent costs.Another advantage of label-free method is the ability to perform quantitative measurement of molecular interaction in real-time, allowing continuous data recording.Moreover, target analytes are detected in their natural form without labeling and chemical modication. The label-free sensing strategies operate through a binding-event-generated perturbation in optical, electrical or mechanical signals.
Optical transducers are widely used due to their high sensitivity with several well established optical phenomena such as surface plasmon changes [7, 8, 9, 10], light scat-tering/adsorption [11, 12, 13, 14]. Most label-free optical biosensors require precise alignment of light coupling to the sensing area, which is a major drawback for point-of-care applications. Therefore, optical sensing can be signicantly improved integrating this approach to severalpassive and active opticalcomponents on the same substrate, allowing the fabrication of multiple sensors on one chip.
Advances in micro- and nanofabrication technologies have facilitated the emergence of micro- and nanoscale mechanicaltransducers capable ofdetecting changes in force, motion, mechanicalproperties and mass that come along with molecular recognition events [15, 16, 17, 18].Among the dierent mechanical biosensors, cantilever and quartz crystal microbalances (QCMs) are the most established techniques.Mechanical bending of a micro- or nanocantilever is monitored as analytes bind, with optical readout typically used to detect the deection or change in stress/strain prole of the cantilever. In one example, Kosaka et al., 2014, [19] developed a sandwich assay that combines mechanical and optoplasmonic transduction which can detect cancer biomarkers in serum at ultra low concentrations.In this array the second antibody is tethered to a gold nanoparticle that acts as a mass and plasmonic label;the two signatures are detected by means of a silicon cantilever that serves as a mechanical resonator for `weighing' the mass of the captured nanoparticles and as an opticalcavity that boosts the plasmonic signal from the nanoparticles, achieving a detection limit of 1Ö10−16 g ml−1 in serum which is at
least seven orders ofmagnitude lower than that achieved in routine clinical practice. This class of biodetectors is very appealing in terms of limit of sensitivity and it is well suited to laboratory applications; however,the application of an external potential to a piezoelectric material (in this case the quartz crystal) produces internal mechanical stresses that induce an oscillating electric eld.Moreover, the resonance frequency shift can be inuenced by many factors, such as changes in mass, viscosity, dielectric constant of the solution and the ionic status of the crystal interface with the buer solution.
The best performances are expected by detectors based on electricalreadout, both in terms of cost reduction and of multiplexing analysis of dierent biomarkers. These sensors integrated with microuidic networks in a Lab-on-a-Chip platform, can develop into easy-to-use, rapid and reliable diagnostic kits to be operated as medical practi-tioner's bench tool [20]. In contrast to the others that are interesting in terms of limit of sensitivity and well suited to laboratory applications but whose cost cannot easily be decreased to make them not available for systematic point-of-care diagnostics.
CHAPTER 1. INTRODUCTION 14 to avoid charge-transport from the electrode to the electrolyte and vice-versa which are inducing a rise of electrochemicalreactions that can change the ionization state of the atoms at the surface.In this way, at the interface, the moving ions in solution are period-ically attracted and repelled from the surface oscillating around a central position and, in rst approximation, they behave like a planar capacitor. This capacitance is called double layer capacitance,CDL, and it is formed by the charged electrode (because of the applied potential) and the layer of mobile ions in the diusive layer within a distance equal to the Debye length1 of the solution.
In 1853 Helmholtz, knowing that at the equilibrium, the charge in the metallic elec-trode is distributed at its surface, modeled the elecelec-trode/electrolyte interface as com-posed of two layers of opposite polarity where the ions in solution counterbalance exactly the charge on the metal and form, together with the electrode, the double layer capac-itance,CDL. After this rst model severalscientists rened the theory but the name
given by Helmoltz to the double layer capacitance remained.
For a mathematical treatment of the interface one can introduce an equivalent circuit described in Figure 1.1, where the two2main components of the equivalent circuit:C
DL
andRchannel are highlighted ([21, 22]).As already explained,CDL models the solid/liquid
interface whereasRchannel idealizes the ionic resistance of the electrolyte solution in the
pool.
In literature can be found many examples of bio-detectors based on Electrochemi-cal Impedance Spectroscopy [1,22, 23, 24]. In general, according to the frequency of
1In electrochemistry the Debye length is the measure of a charge carrier's net electrostatic eect in
solution, and how far those electrostatic eects persist. For a symmetric monovalent salt, is dened as: k−1 =
r
r 0RT
2F2C0 (1.1)
where R is the gas constant, T the absolute temperature, F the Faraday constant, C0 the molar
concentration of the electrolyte, r and 0 are the dielectric constant of the solution and vacuum,
respectively.k−1 gives us information about the electrostatic screening eect of the solution and, as we
can see in Equation 1.1, decreases for increasing ionic strength.
The ionic strength provides the concentration of all ions present in a solution, I = 12Xn
i=1
ciz2i (1.2)
whereci is the molar concentration of ion i (M, mol/L), zi is the charge number of that ion, and the
sum is taken over all ions in the solution.
2considering negligible and therefore not reported the parasitic contribution of the surrounding
Figure 1.1: a) Idealization of the physical system of the measuring setup.b) Equivalent electrical circuit of our biosensor. CDL models the interface electrode/electrolyte while
Rchannel represents the ionic resistance of the channel.
the applied voltage, the capacitance at the electrode/electrolyte interface or the resis-tance of the electrolytic solution is measured [1, 22]. The devices based on resistivity measurements,showed high sensitivities [23, 24] whenever an appropriate choice of the functionalization of the substrate. In resistance immunosensors a redox species is al-ternately oxidized and reduced by the transfer of an electron to and from the metal electrode. Thus, faradaic EIS requires the addition of a redox-active species and DC bias conditions such that it is not depleted [24-1].
Capacitive readout oers the ability to create label-free integrated microsystems with miniaturized dimensions thus saving cost,time and system complexity. The possibility to integrate capacitive biosensor arrays into a single chip along with the enablement of electrical detection are perhaps the most signicant advantages of this kind of sensors. Integration, on one hand,signicantly increases throughput and automation of the de-vice and reduces the size ofthe biosensor as wellas the cost of the diagnostic assays avoiding to consume hundreds of microliters of expensive reagents,while electrical de-tection eases readout electronics design and implementation thus leading to compact portable instruments for point-of-care diagnostics with great results in terms of device sensitivity, e.g. 0.1 pM is the detection limit achieved by Limbut and co-workers[25] in the Bacterial Endotoxin detection using gold electrodes.Qureshi et al.[26], showed a limit of detection of 25 pg/ml for the detection of biomarkers connected to cardiovascular risk by using an assay of interdigitated gold electrodes.
CHAPTER 1. INTRODUCTION 16 the biological receptors are usually specic, interfering substances in the analyte solution may aect both the sensor selectivity and sensitivity[27, 28].False positive or false nega-tive results may be due to the matrix components.A common practice for compensating the eect of interfering parameters is dierential measurements using reference sensors without biological receptors on their surface or with molecules that are expected not to interact with the analyte.
This type of devices may have a great impact in the future of medical diagnostics but, at the same time, the fabrication steps and the functionalization with biomolecules are still very complex.
For all these reasons we decided to concentrate our attention on detectors based on capacitance read-out[29],working thus at low frequencies (which,given the dimensions of our setup and buer conditions, means frequencies lower than 1.3 kHz) where the so called double layer capacitance dominates,as demonstrated by Zou et al.[30]. In this regime, it is crucial to control the voltage applied to the electrodes with high accuracy, to avoid voltage-induce damaging of the functional molecules adsorbed on the gold surface. To this aim, a third reference electrode is often used.In fact, it was demonstrated in a previous work of our group that a three-electrode, miniaturized capacitive device allows for high time stability, enables rapid, real-time response and improves resolution[29], hence being an ideal candidate for fast personalized diagnostics, and/or therapeutic drug monitoring ([31, 32]).
In particular, we concentrated on DNA detection. Oligonucleotides are in fact inter-esting for biomedical and technical reasons:
DNA is the most stable biological
molecule that contains information about the
source organism.This makes it the ideal molecule for pathogen detection and drug discovery.Moreover, the ability to synthesize DNA oligonucleotides allows for the low-cost design of DNA biosensor microarrays;
The unique biomolecular recognition properties of
DNA, based on Watson and
Crick base pairing, can be exploited to detect complementary DNA (or RNA) strands in solution and to bind antibody DNA-conjugates with a high bioanity level;
In DNA sensors,
Figure 1.2:Cross section of the mixed Self assemble Monolayer (SAM) formed by ssDNA-and dsDNA-molecules immobilized on Au(111) gold lm. Because ofthe variation in the persistence length of the DNA-molecules after hybridization (from 1 to 50 nm), the ssDNA is more exible than dsDNA which results in a shorter rod.
dierent, non-(or not fully-)complementary nucleic acids. A single mismatch for a 20 bases long probe can destabilize the complex such that its melting temperature is reduced by 5°C [33].This makes detection using DNA strands extremely specic, given that the probe length is short enough for the mismatch to be signicant, but long enough for the global stability of the molecular complex;
DNA molecules remain intact up to about 80° C.
As temperature is lowered,the
denatured DNA will self-recognize and hybridize once again, so the device can be reusable;
Label-free DNA sensors have been built on many dierent substrates and materi-
als, including gold and carbon, silicon dioxide, diamond, quartz, optical ber and conducting polymers[34].
CHAPTER 1. INTRODUCTION 18 characterization of the device, which means the research of the best operating conditions to perform the various experiments.
It becomes clear that, whether it is for genomic protein application for early screening of diseases,routine walk-in medical checkups,or forensic analysis,controlled design of these sensors is essential.Reliable outcomes of critical medical tests mean less need for repetitions, oering better diagnosis with earlier intervention and more ecient thera-pies. The problem of achieving a repeatable outcome is stillopen. Even in the case of the electrochemical based DNA sensors, the output is tted with an empirical or semi-empirical model, and some conclusions are derived.It is, however,very important to know what mechanisms cause the biosensor to operate.Simulation of the sensors using physical models can shed light into these mechanisms and help answer this question [35].
The aim of the present Ph.D. work is to design, develop and characterize a DNA-biosensor with electrical read-out proving its clinical utility through the detection of circulating cancer biomarkers as microRNAs, in real, complex environment (e.g.blood, serum).
To achieve the aims described above, a miniaturized electrochemical cell was de-signed and developed (cf.Figure1.3). So, starting from this measuring setup, our nal purpose is to implement and microfabricate a biosensing platform based on electro-chemicalimpedance readout with the ultimate goal of performing label-free, real-time measurements of clinically relevant biomarkers.The idea is very simple and consists in functionalizing the gold surface of the WE with surface-immobilized ssDNA-molecules with dierent sequences and perform EIS measurements of oligonucleotides hybridization (DNA, miRNA), in order to exploit the expertise of the group in the functionalization of gold surfaces,or measuring the electrical response ofthe device due to an antigen-antibody biorecognition event, conjugating the protein to an oligonucleotide strand via DNA directed immobilization (DDI) [36, 37].
Then, the optimized device was used to detect the unknown concentration of miRNAs in cell extract and in human plasma of patients with heart failure disease, using derived calibration curves to quantify it. We distinguished two contributions to the variation of dierential capacitance: one, due to the blocking eect of the complex matrix (e.g. human plasma) and a second one, showing a time dependent kinetics, due to the specic miRNA/DNA biorecognition.
Figure 1.3: a) Scheme of the connections of the three-electrode setup. The potential is applied across WE and RE whereas the current is measured across WE and CE.b) Cartoon showing the assembling method of the three-electrode setup.In the foreground, the small silicone pool where the experiments were carried out.The dimensions are not to scale.
Chapter 2
Materials and Methods
2.1 Device Fabrication & Measuring Setup
As already explained in the work of Ianeselli et al.[29], it is possible demonstrate that a device composed of three electrodes guarantees a more stable signal than one with only two electrodes[29,38]. Therefore an electrochemicalcell composed ofthree electrodes was realized:the Working (WE) and the Counter (CE) electrodes are in the micrometer scale while the Reference (RE) is an electrode commonly used in electrochemical cells:a millimeter size (d ' 4 mm) Ag/AgCl electrode obtained from the chlorination of a silver pellet. This kind of electrodes, called second-species reference electrodes, were chosen for their simplicity of manipulating and for their proper working in contact with a solution
containing chloride anions (Nernst equation for these half-reactions depends only with the chloride concentration).
Devices were fabricated using proximity UV-optical lithography based process. Work-ing and counter electrodes were produced on clean microscope slides by a classical lift-o process.
First of all, the clean1 slides were dehydrated at 200 °C for 5 min and then treated
with O2-plasma in order to increase the adhesion of the photoresist. The parameters that we used to increase the hydrophilicity of microscope slides are listed in Table 2.1.
1The cleaning protocol consists in three successive baths in acetone, 2-Propanol and soapy milli-RO
water in a sonicator water bath for 45 minutes in total, and thus rinsed with milliQ water.
Element Power (W) Flux(sccm) P (mbar) Bias (V) Exp. time (min)
Oxygen O2 60 30 1.5 × 10−1 ' 300 5
Table 2.1: Parameters for the oxygen plasma that we set in order to increase the adhesion of resist on glass substrate.The plasma treatment was performed in the chamber of the machine for reactive ion etching (RIE).
Successively the slides are rapidly cooled down in a nitrogen stream and immediately spin-coated with a UV-photo-sensible resist. In this thesis work we employed mostly MEGAPOSITTM SPRTM 220 1.2 (Series Photo-Resist) as a positive resist.The type of the resist, negative or positive (as here),denes how the mask image is transferred. For negative resist the exposed areas remain on the slides after development whereas for positive resist the exposed areas are removed during development becoming soluble to the developer solution. According to the fabrication needs,one can vary the lm thickness
CHAPTER 2. MATERIALS AND METHODS 22
Resist SPRTM 220 1.2 SU8-2002
Dehyd. glass slides 200°C for 5 min 100° for 5 min
Spinning 500 rpm for 5s 3000 rpm for 35 s
2500 rpm for 25 s Soft-Bake
95°C for 1 min 65°C for 1 min
thermal ramp 95°C→ 115°C thermal ramp 65°C→ 95°C
115°C for 90 s 95°C for 2 min
Exposure 13 sec @ 3 mW/cm2 9 sec @ 5 mW/cm2
Post-Bake No
65°C for 1 min thermal ramp 65°C→ 95°C
95°C for 1 min
Development ∼ 28 sec in MF24A ∼ 40 sec in SU8-Dev.
Rinse in milli RO water in IsoPropanol
Hard-Bake 115°C for 2 min No
Table 2.2: Lithographic steps to be performed for the positive photoresist SPRTM 220 1.2.
After the described photolitographic steps, the so patterned slides were metalized using an e-beam evaporator.First the micro-patterned microscope slides were inserted in the evaporator and two metal layers, 20 nm of Titanium (in order to facilitate the adhesion on glass,) followed by an 80-nm Au layer,were deposited under high-vacuum (' 3 × 10 −6 torr) in order to prevent oxidation and obtain smooth and uniform layers.
Following evaporation, the slides underwent lift-o where the inverse patter layer is sacriced. In this process the samples were immersed in an acetone bath overnight and rinsed thoroughly with acetone to remove the last remnants of sacricial covered resist and eventually cleaned with isopropanol to remove any halo left by acetone bath. After lift-o process the samples with the two metalized electrodes required another complete lithographic sequence in order to dene the area of the working and counter electrode in contact with the solution and isolate the rest.In this last part of fabrication protocol, electrodes were coated with an insulating layer (about 3µm) of NANOTM SU8-2002 (negative resist),shaped to expose only the circular part of the WE and the CE performing an aligned lithography using the mask aligner MJB3 by Karl Suss, Germany.
Pictures of the slides patterned with the micro-electrodes,are shown in Figure 2.1. The microfabricated electrodes are composed by two gold pads for wire-connection.The WE has a diameter ofDWE = 100 µm (⇒ A WE ' 7.85 × 10−5 cm2), and is connected to
Figure 2.1: a) Picture of the microfabricated WE and CE for the three-electrode setup. b) SEM image of the microfabricated working (WE) and counter (CE) gold electrodes. The gold electrodes appear lighter than the background of the picture.c) The patterned insulation layer is shown and it is higher with respect to the plane of the electrodes. EHT=2 kV and Mag 82x.
of 300µm that encloses the WE. A layer of SU8 2002 with a thickness of 2.9µm (rec-ognizable in Figure 2.1) was realized in order to reduce the active area of the electrodes (pac-man shaped in Figure 2.1b) ) and insulate the electrical path that connect WE and RE with the connection-pads.
In order to connect the electrodes with the electrical instruments, a sample holder suitable to the features of our patterned slides has been developed.The sample holder is shown in Figure 2.2a). From the picture, the micro-electrodes have been constrained between two Plexiglas slides.The measurements were carried out in a small silicone pool with a diameter of 6 mm and a height of 4 mm, holding a 125µL volume. In this way the reference Ag/AgCl pellet electrode (diameter,d = 5 mm), could be inserted directly in the solution through the hole in the top Plexiglas slide and placed just above the microelectrodes (see Figure 2.2).The electrical signal was collected from the gold pads (4 × 4 mm2) by a circuit board with a specic design equipped with SMA connectors
and spring-loaded pins (see Figure 2.2). In order to shield the device from parasitic contribution of the surrounding environment, we further designed and realized a very robust Faraday cage (see Figure 2.3a) and b)).The cage was made of solid 1 cm thick aluminum plates and with a total weight of about 10 kg. In this way also mechanical vibrations could be eectively shielded and measurements at low frequencies became easier.
CHAPTER 2. MATERIALS AND METHODS 24
Figure 2.2: a) Sample holder for our miniaturized biosensor.The microscope slide with the micro-electrodes are constrained between two Plexiglas slides.The measurements were carried out in a small pool with a diameter of 6 mm and a height of 4 mm, holding a 125µL volume. The reference Ag/AgCl pellet electrode is immersed directly in the solution through the hole in the top plexiglas slide and placed just above the microelec-trodes. The electrical signal is collected from the gold pads (4x4 mm2) by a circuit board of custom design with SMA connectors and spring-loaded pins.
(programmable via Potmaster Software) shown in Figure 2.3c). The advantage of the bipotentiostat, was the possibility to handle three electrodes.The third electrode in fact, which acts as Reference Electrode (RE), made of Ag/AgCl is used for the high stability of its potential in contact with the electrolytic components of human blood, to assure a ne control of the absolute value of the potential applied across WE and RE and thus the measurements of the current owing between WE and CE.
The potential applied during the measurements was an AC sinusoidal voltage center at 0 V, whose amplitude was optimized to the needs of the specic experiment and usually ranged from 10 mV to 150 mV. The stimulus signal for each time was mediated on four frequencies:100 Hz, 200 Hz, 250 Hz and 400 Hz.These frequencies were chosen because in this regime the system has a purely capacitive response, (see Figure 2.3d)). At each frequency we collected 200 complete periods from which we computed the root mean squared value ofthe measured current,Irms = 2πf · V rms · Cd, and the relative
uncertainties using error propagation analysis.
CHAPTER 2. MATERIALS AND METHODS 26 (500 ms per each of the four frequencies).
In conclusion, by calculating the dierential capacitanceCdfrom the currentIrms, it
can be possible to derive information about the DNA-SAM composition and assembling. As already explained in the work of Ianeselli et al.[29],in fact, our device measures the dierential capacitance Cd at the electrode/electrolyte interface that, in our
cong-uration, can be set equal toCDL, the double layer capacitance already introduced in the previous section, and dened as:
CDL ' C d= ϑσϑϕM (2.1)
Cd reects the charge density (σM) change at the metal surface for a small variation
of the applied potential (ϕ) between electrode and solution. In the case of a biofunc-tionalized metal electrode immersed in a saline solution,Cd can be modeled as a series
connection of two capacitances ([39, 40]):the capacitance due to the absorbed layer of molecules (Cmol in Figure 2.4a)) and the one related to the ions in solution, the so-called
diusive layer capacitance (Cions in Figure 2.4a)).
According to this model, proposed by Helmholtz in 1853, the molecules in solutions can either lay down on the surface or stay in the diusive layer according to Boltzmann's distribution at a distance from the surface equal to the Debye length of the solution.
Normally, Cions has densities of the order of 40 µF/cm2, larger than Cmol(≈ 10
µF/cm2, in agreement with literature [41]). The nal capacitance is the inverse of the
sum of the inverse of the two linked capacitors and thus the dominant contribution is due to the variations ofCmol that raised upon molecular adsorption on the electrode
surface including height changes, changes in the electrical charge density and substitution of water molecules in the biological layer since the SAM layer is composed of DNA
monolayer and of the ions solvating the strands (Figure 2.4b)).
The resulting capacitor can be treated as a planar capacitor and thus:
Cd = 0 r Ad = 0 r AWEd (2.2)
where r is the dielectric constant of the molecular layer and0 the vacuum permit-tivity, A is the surface occupied by the layer (here is equal to the area of the gold working electrode,AWE ' 7.85 × 10−5 cm2) andd is the thickness of the monolayer.From
Figure 2.4: Idealization of the electrode/electrolyte interface.a) The rst layer consists of ssDNA on Au(111) surface. The counterions in solution, just above the molecular layer, composes the second layer modeled with a second capacitance,Cions, in series
with the rst one. b) The total capacitance at the interface is the sum of capacitances linked in parallel and associated to the area composed of DNA strands or ions.
although it represents a good approximation of our device, , Cd is independent of the applied voltage, and this is the main reason that prompted me to look for a model that took account the applied potential and also shows the dependence of andd from the target concentration in solution.
2.2 Oligonucleotides
CHAPTER 2. MATERIALS AND METHODS 28
Figure 2.5: The structure of the DNA double helix. The red-colored backbone is com-posed of sugar-phosphate bonds and the bases in the structure are color-coded by element and the detailed structure of two base pairs are shown in the right. Image courtesy of
Serena Rosa Alfarano.
Greek "oligo" which means few or small) are short DNA or RNA molecules characterized by a sequence of nucleotide residues that make up the entire molecule and constitute the keystone of the research described in this thesis work.
A nucleotide is made of a nucleobase (also termed a nitrogenous base), a ve-carbon sugar (either ribose or 2-deoxyribose depending on if it is DNA or RNA) and one phos-phate groups.
Nitrogenous bases are typically classied as the derivatives of two parent compounds, pyrimidine (include uracil, thymine, cytosine) and purine (adenine and guanine), abbreviated as U, T, C, A, and G, respectively.Uracil and thymine are identical except that uracil lacks the 5' methyl group.
Most of DNA exists as double stranded DNA, consisting of two oriented, comple-mentary, polynucleotide sequences.Double helix is2.2 ÷ 2.6nm large and one base-pair corresponds to approximately0.34nm of length along the strand.
From Figure 2.5 one can see that lateral DNA structure (backbone) is composed by repeated phosphate groups between two sugars with ve carbon atoms (deoxyribose for DNA and ribose for RNA): each phosphate group is bound to the30carbon of the rst
sugar and the 50 carbon of the following sugar. In this way DNA is oriented in the
this combination ts the constant width geometry of the DNA spiral. TheG ≡ C pair is bound by three hydrogen bonds,while A = T bound by two hydrogen bonds,which means that DNA sequence with a high GC-content has a greater thermal stability than DNA with a lower percentage of GC bonds.
The rst interesting (for ours studies) property of DNA that deserves to be empha-sized is its ability to create a self-assembled monolayer on (and not only) gold surfaces. Among the many SAM system investigated, those made by adsorbing alkanethiols on single crystal surfaces have been more frequently studied and, chemically modifying the end of a DNA single strand with a thiol group, a Self Assembled Monolayer of DNA on a gold surface where thioloxidizes and bends to the single gold crystal,can be formed where the strands, inside the monolayer, can interact through electrical and Van der Waals interactions.
Another, most relevant feature for this work is that DNA is a highly charge poly-electrolyte. As already stated, DNA is basically a polymer of nucleotides which are held together by covalent bonds formed between the phosphate groups, each of which forms an ester with a hydroxyl group of the pentose of the next nucleotide.This involves two of the three OH groups of the acid, leaving the last one free to ionize. This ionization leaves a negative charge on each phosphate group [42].Such molecular structures can be used to detect complementary DNA (or RNA) strands in solution and to bind antibody DNA-conjugates in bioanity assays for novel applications, ranging from the study of protein networks to monitor the progress of diseases.
2.3 An insight on SAM formation
Self-assembly forms the basis for many natural processes including protein folding, DNA transcribing and hybridization, and the formation of cell membranes. The process of selfassembly in nature is governed by inter- and intra-molecular forces that drive the molecules into a stable, low energy state.These forces include hydrogen bonding, elec-trostatic interactions, hydrophobic interactions, and van der Waals forces.
CHAPTER 2. MATERIALS AND METHODS 30 mechanism of the self-assembly process has been well studied and elucidated.Strong and Whitesides[43],via electron diraction studies of high density monolayers,have found that a typical alkanethiol monolayer forms a ( √3 ×√3 R30°) structure on Au(111) with the thiol chains tilted approximately 30 degrees from the surface normal.
As with self-assembly in nature, there are several driving forces for the assembly of thiols onto noble metal surfaces. The rst is the anity of sulfur for the gold sur-face. Dubois et al., have found that the sulfur-gold interaction is on the order of 45 kcal/mol[44], forming a stable, semi-covalent bond,in comparison, the C − C bond strength is∼ 83 kcal/mol.
The next driving force for assembly is the van der Waals interactions between the methylene carbons on the alkane chains. For alkanethiol monolayers,this interaction causes the thiol chains to tilt in order to maximize the interaction between the chains and lower the overall surface energy[45].
An alkanethiol is compound containing an alkyl group (−CxH2x+1) joined to a
mer-capto group (−SH ). The substrate here discussed is Au(111). Each alkanethiol chain can be divided into three parts:
headgroup (linking group) which guides the self-assembly process on each type
of substrate through a strong bond;
the backbone (main chain) that interacts with other chain backbones via van der
Waals and hydrophobic forces.This ensure an ecient packing of the monolayer and contribute to stabilize the structures;
the specic terminal active group which confers specic properties to the sur-
face and is able to bind dierent molecules by weak interactions or covalent bonds. In the case of SAM-alkanethiol, the reaction may be considered formally as an oxidative addition of the S-H bond to the gold surface, followed by a reductive elimination of the hydrogen.
From Figure 2.6a), we can follow the steps of how a SAM composed of alkanethiol chains on Au(111) forms as a function of deposition time[46]:
2. Lying down phase. After physisorption, thiol molecules chemisorb on the Au(111) substrate through the sulfur headgroup in a process that takes at least some min-utes. During the process the thiol molecule loses the mercaptan H atom,trans-forming itself in a thiolate. This kind of nucleation happens easier at Au step edges and forms islands composed of adsorbed chains.The phase is called "lying down" because the axis of the chain is parallel to the substrate;
3. Nucleation of standing up phase. After nucleation, the islands grow and there is the increase the surface coverage ofthiolate species on the Au surface. "Standing up" is referred to molecules which have their axis perpendicular to the substrate. This phase is governed by the balance between intermolecular and molecule/substrate interactions and the gold surface response to the chemisorp-tion process.The competing forces that determine the SAM ordered structure are the interaction between the headgroup and the substrate,which involves a large chemisorption energy (30 kcal/mol), and the interchain van der Waals forces.Al-though van der Waals interactions are weak with respect to chemisorption ones, they inuence and stabilize molecular self-assembly.
4. Standing up phase. The self-assembly takes place in two consecutive nucleation and growth processes.The standing up is the second one and it leads to rotated domains of lying down molecules, and later to domains of standing up molecules, irrespective of the environment used for SAM preparation.
The self-assembly takes place in two consecutive nucleation and growth processes that lead to rotated domains of lying down molecules, and later to domains of standing up molecules, irrespective of the environment used for SAM preparation.
From the alkanethiol SAMs the extension to DNA-SAMs was easy:chemically mod-ifying the end of a DNA strand with a thiol group permits to form a Self Assembled Monolayer of DNA on a metal surface where thiol oxidizes and bends to the metal, with the strands interacting through electrical and Van der Waals interactions. The main dierence is that DNA is negatively charged and, in order to screen the negative charge of DNA backbones, a saline buer solution is usually used..In fact positive ions of saline solution shield the DNA chains electrostatic repulsion and allow the SAM to form.
CHAPTER 2. MATERIALS AND METHODS 32
Figure 2.6: a) Scheme of the dierent steps taking place during the self-assembly of alka-nethiol on Au(111): (i) physisorption, (ii) lying down phase formation, (iii) nucleation of the standing up phase, (iv) completion of the standing up phase[46].b) commensurate √
3 × √3 R30° structure on Au(111). Blue lines are primitive vectors of the substrate and red lines are the primitive vectors of the adsorbate lattice that are larger and rotated with respect to the blue ones.Image courtesy of Serena Alfarano.
(cross-linking bases), which means that DNA can physiadsorbed to the gold surface via N-Au bonds (see Figure 2.7 for a representation scheme).This last case of physisorption bonding, usually occurs at energy of 6 kcal/ mol, at variance with a chemiadsorption bonding S-Au at higher energy (30 kcal/ mol)[45].
2.4 ssDNA-SAM Density
The density of our ssDNA-SAM in solution can be controlled varying the incubation time, that is the time that the solution, containing a given concentration (1 µM in our case) of thiolated DNA molecules, is left in contact with the Au(111) surface of the WE. Obviously, increasing the incubation time more dense will be the DNA-SAM layer. Principally we used in this work two dierent incubation times: low density SAM (LD-SAM) realized keeping the samples in contact with the functionalizing solution for 10/15 min whereas high density SAM (HD-SAM) formed increasing the time to 70 min.
Figure 2.8: Georgiadis et al.results[48]:ssDNA and dsDNA exponential density behav-ior as function of incubation time.
CHAPTER 2. MATERIALS AND METHODS 34 (due to the electrostatic repulsion between DNA strands) whereas, higher ionic strength solutions are associated to higher probe coverage.Moreover, at a lower ionic strength adsorption is slower in the rst few hours than at higher ionic strength. This is due to electrostatic phenomena connected to the Debye's length of the solution.
In order to quantify the density of ssDNA layer autoassembled on the surface of WE, XPS measurements were carried out.The idea was based on the experiments of Tarlov and co-workers[49] and the density results were compared with those obtained by Georgiadis et al.[48]. Dierent incubation times has been considered: 5, 12, 20 and 30 minutes, 1, 3, 5 and 24 hours.
The analysis were performed onto nitrogen spectra N1s.They are univocally related to the presence of DNA on Au(111): so spectra were taken in the binding energy range 390-410 eV, because N1s peak is centered at 395 eV. The 24 hours incubation was selected as reference density and it is associated to the maximum possible density for ssDNA and all spectra are normalized with respect to it.In fact, at suciently high probe densities,
DNA is expected to be in the strongly charged regime of polyelectrolyte brushes ([50, 51]). The electrodes were post-treated for 1 h in contact with a solution of TE NaCl 1 M containing mercaptohexanol (MCH) 1 mM. MCH-molecules chemisorbed onto the surface through S-Au bond ensuring the displacing of aspecic ssDNA bonds.We have chosen to work in this way because the XPS data analysis modelcan be simplied by excluding those nitrogen bases of the molecules laying down on the surface which will not participate in DNA-hybridization processes.
The ssDNA-SAM probe density was calculated assuming an hexagonalpackage for DNA onto Au(111) (see Figure 2.9):DNA is treated as a solid and rigid chain of diameter 1nm, it can be done because the high density structure forces DNA strands to stand up from the surface.Steric interaction are taken in account assuming that DNA strands in HD phase locate themselves at a minimum distance of 2nm[52].
The measurements were performed using a conventional Al/Kα source with an elec-tron pass energy of 20 eV. The spectra were collected as the sum of following scan on an energy window of 20eV and sampled with a step of 0.1 eV and with a time of 500 ms. For each sample the Fermi level (that should be positioned at binding-energy 0 eV) was taken and all spectra were thus rescaled with respect to it.
Figure 2.9: a) Top view of hexagonalpackage ofssDNA molecules (cylinder) on gold surface.b) ssDNA molecules (solid blue cylinder) are large 1nm and because of coulomb electrostatic repulsion they are separated by 2nm distance (transparent blue dotted cylinders). Image courtesy of Serena Rosa Alfarano.
and the nal density evaluated for each incubation time, respectively.
CHAPTER 2. MATERIALS AND METHODS 36
Figure 2.11: ssDNA density as function of incubation time.Results are comparable with those obtained by Georgiadis et al.[48].
2.5 Experimental procedures
DNA functionalization of the gold electrodes was carried out using the well-established procedures for single stranded (ss) DNA2 self-assembled monolayers on gold ([37,48]).
The electrodes were wetted for 12 min, with a drop of 1 µM thiolated (C6) ssDNA (HS − (CH 2)6-5'-ctt atc gct tta tga ccg gac c-3', called F5-SH ) in a high-ionic-strength
buer TE NaCl 1 M at pH 8. In this way a low density ssDNA SAM ( 2.1 ± 0.4 × 1012 molecules/cm2) was obtained (more details can be found in Section 2.4 and accord-ing to[48]). After SAM formation, in order to remove aspecically bound DNA-molecules, the electrodes were thoroughly rinsed with the same buer solution, at phys-iological concentration, used for the measurements,e.g. KCl 100 mM, PBS 1x, etc. Then the dierential capacitance at the electrode-electrolyte interface was measured. Hybridization was performed by wetting the functionalized working electrodes with a drop of the same buer solution containing the complementary DNA strand (5'-ggt ccg gtc ata aag cga taa g-3', called cF5 ), in dierent concentrations.The used 44-mer ssDNA oligonucleotide has a sequence (HS − (CH 2)6-5'-caa aac agc agc aat cca aag atc aga cac
ccg att aca aat gc-3', called dpnII_SH ).
We chose to work at low surface probe densities because,in this way, electrostatic and steric eects are minimized and a complete hybridization and relatively fast kinetics can be achieved. Instead, in a high density regime, experimental measurements have revealed a strong suppression of DNA target hybridization ([37, 48, 53]).Moreover, we noted that a further passivation treatment of the surface of the electrode with molecules as MCH-molecules, prevented the detection of the DNA-hybridization masking the signal and thus reduced the sensitivity of the device.
In the DNA/miRNA hybridization experiments performed here, the DNA sequence (HS − (CH 2)6-5'-cga agg caa cac gga taa cct a-3' ) was chosen to be complementary to
a miRNA target namely hsa-miR-154-5p (5'-uag guu auc cgu guu gcc uuc g -3' ), up-regulated in heart failure ([54, 55]). In addition, a murine miRNA, mmu-miR-351-5p
(5'-cag gct caa agg gct cct cag gga-3' ), was used as negative control and complementary to the ssDNA sequence: mmu-miR-351-5p-comp-SH ( HS − (CH 2)6-5'-cag gct caa
agg gct cct cag gga-3' ).In these experiments the miRNAs detection was performed in buer solutions (PBS 1X, KCl 100mM) and human extract and then implemented in human plasma samples having dierent levels of heart failure.
All the used oligonucleotide-molecules were purchased from Biomers.net. A complete list of oligonucleotide sequences and nomenclature can be found in Figure 10 in Appendix D.
2.5.1
De-hybridization protocols for device regeneration
Thermal de-hybridization cycles were introduced to test the reusability of the device. The thermal treatment in the case of DNA/DNA hybridized SAM on the WE, in buer solution, consisted of sample incubation in a basic solution (pH = 9 ) of TE buer for 1 hour in oven at a temperature 10° C higher than the melting temperature of the specic DNA sequence (e.g.55°C for the sequence F5 previously introduced).
Chapter 3
A Biosensor for Direct Detection of
DNA Sequences
As already mentioned in the introduction, the nal goal of this work is to produce a portable capacitive device to be used as a point-of-care medical diagnostic tool. To-wards this goal, the unique biomolecular recognition properties of DNA are exploited to detect genomic biomarker with high anity. To optimize the sensitivity of the device we investigated DNA-SAM electricalproperties using dierent physiological conditions mimicking the complexity of blood composition. After ssDNA-SAM characterization, the hybridization process has been characterized,in the case of fully- and partially-matching target DNA sequences.
3.1 DNA-SAM Electrical Characterization
As already stated in Section2.1 Electrochemical Impedance Spectroscopy (EIS) mea-surements were performed on a three-electrode conguration to characterize the DNA SAM.
The WE and CE are microfabricated gold electrodes, immersed in an electrochemical pool, whereas the RE is a classical mm-sized Ag/AgCl pellet electrode accessing the experimental pool from the top. As in a normal electrochemicalsetup, the potential is applied across WE and RE while the current, owing in the experimental pool, is measured across WE and CE. In order to compare our measurements with standard conditions in electrochemical setups,the RE was calibrated with respect to a standard
Ag/AgCl RE (sRE) placed in a solution of saturated KCl 3 M. Practically, we measured the open circuit potential (OCP1) of our RE with respect to the sRE (∆V
Re−sRE ) in the
dierent working buer solutions used in this thesis-work. In this way, we can relate the potentials measured in our setup to the sRE (∆VWE−sRE ) according to the following
equation:
∆VWE−sRE = ∆VWE−RE + ∆VRE−sRE (3.1)
where∆VWE−RE is the bias potential measured between WE and RE. In this way we
derive the correcting parameter to set in the bipotentiostat control software to assure a ne control of the applied potential during the whole experiment and to allow a direct comparison with other data available in literature.
After this calibration, we tested the eect of the applied potential (details in section 3.1.1) and of the buer conditions (see section 3.1.2) on the measurements of DNA-SAM capacitance. To optimize the conditions for best DNA hybridization, dierent SAM densities were exploited.
3.1.1
The eects of the Applied Potential on the ssDNA-SAMs
Prior to SAM-formation we treated the gold surface with plasma etching which makes the surface at and smooth (0.5nm roughness, as measured with Atomic Force Microscopy). Then, we prepared the DNA-SAm as described in section 2.5, using the sequence F5-SH. Capacitance measurements as a function of the applied potential were performed in KCl 100 mM buer solution, as described in section 2.1.
In Figure 3.1a) we show, as an example, the dierential capacitance measured for a low density (LD) ssDNA-SAM obtained by incubated 1 µM of ssDNA for 12 minutes and applying a bias potential ranging from 10 mV to 100 mV. Cd has been measured for more than a hour, to prove the stability of the system.
From the solid red line in Figure3.1b), we noticed a sharp increase of the capacitance, raising the applied potential from 25 mV to 100 mV. This sharp change of Cd can be
related to thiols desorption which led us to terminate the measure in order to avoid SAM damaging. At this point, EIS measurements with dierent applied potential at dierent densities were carried out.Figure 3.1b) show the dierential capacitance measured for
CHAPTER 3. A BIOSENSOR FOR DIRECT DETECTION OF DNA SEQUENCES40 six dierent ssDNA-SAM densities obtained with six dierent incubation times: 12, 15, 20, 30, 45 and 70 minutes, as described in Section 2.4.
Figure 3.1: a) Dierential capacitance measured for a low density (τinc = 12 min)
ssDNA-SAM in KCl 100 mM at four dierent amplitudes for the applied potential: 10, 50, 75 and 100 mV. b) Dierential capacitance, Cd vs the amplitude of applied potential
(Vapp) in KCl 100mM at six dierent incubation densities, from the top, density with
12 minutes of incubation, then 15 minutes, 20 minutes,30 minutes, 45 minutes and 70 minutes. Increasing the SAM-density we can observe a decrease of Cd; furthermore, in
each case, for larger applied potential the measured capacitance is larger.
that the system is stable vs time at each of the applied potential. In Figure3.1b) we checked further the eect of the potential on SAM of dierential densities. The curves of 3.1a) correspond to the same LD ssDNA-SAM of the top red curve of 3.1b). The lowest two curves (dark green and orange markers) correspond to a high density (HD) DNA-SAM, obtained through a DNA incubation of 70 minutes. While HD SAM can support applied potential as high as 400 mV,for LD SAM a potential higher than 100 mV is already inducing a slow desorption of SAM molecules, as can be seen from parallel AFM measurements. A dependence of the DNA-SAM stability to the applied bias on the SAM density is not new in literature: LD SAMs have a poor degree of order, are not stabilized by intermolecular interactions, and their desorption eciency is higher[56, 57]. The monotonic decrease of the dierential capacitance by increasing the ssDNA-SAM density observed in Figure 3.1b), can be explained through the planar approximation (cf. Eq.2.2) used to describe Cd. According to this equation and sinceAWE remained
constant for all devices,Cd variations depend mainly on changes in the height of the
molecular layer,d, and variations in the dielectric constant. As already said, LD DNA-SAMs are less packed ([35, 37, 47, 48]) and much less ordered than HD DNA-DNA-SAMs.This corresponds to a higher free volume of individual ssDNA molecules, pinned to the gold surface through the thiol linker. As a consequence,the corresponding average height of the SAM is lower, the lower is the density. In turn a lower DNA density allows a greater number of water molecules to penetrate the monolayer,contributing more and more to the total dielectric constant of the area close to the electrode,probed by our measurements.The cartoon in Figure 2.8 provides a graphical representation of the LD and HD-SAMs, and of the water molecules in between.
The eect of C d increase with the applied potential, at a given DNA-SAM density, is instead due to polarization eects. The bias potential polarizes the hydrated ions[58, 59] within the layer closed to the electrode surface increasing the dielectric constant. To prove the role of the hydrated ions polarization in Cd measurements,the complex
CHAPTER 3. A BIOSENSOR FOR DIRECT DETECTION OF DNA SEQUENCES42
Figure 3.2: Bare electrodes measurements carried out in KCl100 mM: a) Dierential capacitance as function ofthe applied potential. b) KCl-doped aqueous solution's di-electric constants showing their trend with a more and more high applied potential.
Figure 3.3: a) Dierential capacitance, C d vs the amplitude of applied potential (Vapp)
in MgCl233mM at four dierent incubation time, going down from the red curve at the
top of the graph, density with 12 minutes of incubation, 20 minutes, 45 minutes and 70 minutes. The slow increasing of the capacitance as function of the applied potential for each SAM density is due to the double charge of the ions which screens the eect of the potential. b) ssDNA with 44bases.Total capacitance as function of the applied dierent potential in a buer solution of KCl 100mM. Dierent densities are explored. From the top: density with 15 minutes of incubation, then 45 minutes and 70 minutes. In each case increasing the applied potential, a higher capacitance is measured.Lines are guides for the eye
3.1.2
Divalent salt eects
Experiments similar to the ones described in the previous paragraph were carried out in the case of a divalent salt MgCl2. To get a direct comparison with KCl 100mM, we
chose kept the ionic strength the same, choosing a solution of MgCl2 33 mM in order to
have the same ionic strength.
CHAPTER 3. A BIOSENSOR FOR DIRECT DETECTION OF DNA SEQUENCES44 ssDNA monolayer in the case of MgCl2 solutions [62, 63].As a consequence, we observe
a slower increasing of Cd as function of the applied potential for each SAM density due
to a more eective screening the DNA backbone.
3.2 The biophysics of DNA hybridization
After the electrical optimization of the ssDNA-SAM and of the read-out conditions as explained in the previous section3.1,we concentrated on study of the kinetics of DNA hybridization. The capability of detecting DNA hybridization in real time and in-situ is of critical importance for applied genomics,drug discovery,gene expression proling and other applications.These studies are based on the detection of interactions between oligonucleotides immobilized on the solid support, or probes, with analyte target nucleic acids present in solution.Binding, or hybridization, between probes and targets to form an immobilized duplex depends on the degree of complementarity between the probe and target base sequences and on the steric availability of the surface-immobilized probes.
Many techniques have been employed in order to study DNA-hybridization [64,65, 66], including EIS-experiments [67]but never in-situ, i.e. following the process while occurring, nor in capacitive conditions. As we will show in this chapter, monitoring Cd as a function of hybridization time will give us additional information to be used in DNA-detection.
To detect DNA hybridization, we proceeded as follows: the WE was functional-ized with a low-density ssDNA adopting the procedure described in Section 2.5. The averageCd value measured in 100 mM KCl saline buer at 10 mV applied potential
was(1.06 ± 0.01)nF. An example of this type of measurement is shown in Figure3.4, red curve. From the measuredCd, we estimated a capacitance density at the working
electrode of≈ 10 µF/cm2, in agreement with the literature[41] and with the theoretical
value extracted from eq.2.2.In fact, assuming a SAM-height of2.38±0.07, as determined via AFM measurements [38], and using the ssDNA dielectric constant values found in literature ([68, 69, 70, 71]), a Cd variation between 0.6 nF ( ssDNA ≈ 20 ) and 2.3 nF
As a result, the height depends on the SAM density (ρSAM ) and salt concentration (C0) according to: h ∼ ρSAMC 0 1/3 ' 2.18 ± 0.41 nm (3.2)
Attested the stability of the signal, the ssDNA SAM hybridization kinetics was then followed in-situ, in real-time. The Cd variation due to strand pairing is represented
by the blue curve of Figure3.4, for the case of 10 pM complementary DNA in same salt solution. To follow hybridization kinetics, we initially measured at a rate of 4 measurements/min for 15 min, then we slowed down to 1 measurement/min untill theCd
dierential variations between successive points were lower than 6% which we considered to be the steady-state of our measure. The two dierent sampling steps are due to the need to highlight the part of the hybridization process more specically connected to the direct channel of DNA-hybridization (target approaching directly from solution, without diusing along the surface. More details can be found in Chapter 4).
The Cd measured at steady-state was(0.918 ± 0.001)nF for this concentration of
complementary DNA, which correspond to a decrease in capacitance of13.4% upon DNA hybridization. The lowering of the capacitance upon hybridization can be explained by the height increase of the SAM upon hybridization, due to the dierent persistence length2 of ss- and dsDNA, and by the replacement of water molecules (with a high
dielectric constant, H2O ' 76.7 ± 0.2, for 100 mM KCl-doped aqueous solution[72]) with
DNA strands (lower ) upon DNA pairing.
After this rst proof of principle, we proceeded with DNA-hybridization in real time. As a rst step, the device was calibrated for DNA/DNA hybridization in the same experimental conditions exploring the dynamic range of detectable complementary DNA concentration over 6 orders of magnitudes, from 1 pM to 100 nM. The reusability of our three-electrode sensor obtained by using the thermalregeneration procedure described in Section 2.5.1, is shown in 3.5.
At each cycle, the Cd value after thermal regeneration recovers the initialCssDNA
value with an error ranging from 1% to 3% (Table 3.1) attesting that the used protocol does not damage the ssDNA probe layer on the electrode, despite of the long measuring
2Dened as the length beyond which the polymer direction is becoming random. From literature
CHAPTER 3. A BIOSENSOR FOR DIRECT DETECTION OF DNA SEQUENCES46
Figure 3.4: Kinetics of DNA-hybridization via dierential capacitance measurements. The red signal represents the dierential capacitance of a low density ssDNA SAM func-tionalized WE measured in KCl 100 mM and applying 10 mV of bias.The hybridization (blue points) was per-formed by adding in the experimental pool 10 pM complementary DNA. Cartoon-insets show the variation of thickness and dielec-tric constant between a layer of ss- (on the left side) and dsDNA (on the right side), respectively.The green and orange-dotted lines represent the ts based on rst-order Langmuir adsorption kinetics and double exponential kinetics, respectively.
time, approximately 20 hours. Hybridization was then performed after each regenera-tion cycle, using dierent concentraregenera-tions of complementary-strand DNA (from 1 pM to 100 nM, see Figure 3.5),monitoring the dierential capacitance at the WE versus the incubation time. At each hybridization process, the measurements were run until theCd
dierential variations between successive points were lower than 6%,as just explained. Such steady-state Cd values,namedCdsDNA , are shown in Table 3.1. These
equilib-rium Cd-values are monotonically inversely proportionalto the used concentration of
DNA target. This can be rationalized considering that the capacitance decrement is proportional to the number of hybridized sites at the electrode as can be described by the following adsorption model[73]:
θ (t) = θmax (kon· [cDN A]) a
ka
of f + (kon · [cDN A])a (3.3)
Figure 3.5: Hybridization and de-hybridization cycles on the same low density ssDNA SAM in KCl 100 mM. The [cDNA] concentration increases for every prole by one order of magnitude going from the blue curve (1 pM) to the right of the graph, mauve curve (100 nM). In the time scale (x-axis), it was considered the time delay necessary for the thermal regeneration of the electrodes.
desorption rate constant, respectively.
[cDNA] C(nF)ssDNA ∆(%)ssDNA C(nF)dsDNA
1 pM 1.07 ± 0.01 1.7 0.93 ± 0.01 10 pM 1.08 ± 0.01 2.5 0.92 ± 0.02 100 pM 1.02 ± 0.02 2.9 0.89 ± 0.01 1 nM 1.05 ± 0.01 0.83 ± 0.01 10 nM 1.06 ± 0.01 0.4 0.69 ± 0.01 100 nM 1.06 ± 0.01 0.4 0.61 ± 0.01
Table 3.1: Dierential capacitance measured at the ssDNA functionalized WE before (Cd−ssDNA ) and after hybridization process, Cd−dsDNA . The experiments were carried
out on the same regenerated SAM-probe lm for all the six dierent [cDNA] shown in Figure 3.5.
CHAPTER 3. A BIOSENSOR FOR DIRECT DETECTION OF DNA SEQUENCES48 isotherm where the degree of heterogeneity (distribution of binding energies) increases as the value of a decreases.
Figure 3.6: Calibration curve of the three-electrode device. The response is shown as percentage capacitance variation upon hybridization of the ssDNA-SAM WE, as a func-tion of the complementary DNA concentrafunc-tion in solufunc-tion (KCl 100 mM) and tted with a Hill equation (dotted black curve). The solid red line represents the average variation of Cd when a non-complementary DNA is present in solution at high concentration (1
µM), whereas the dashed red lines represent 3 standard deviations around this value. In addition, the maximum variation of Cd upon hybridization,Cd% (calculated from
CssDNA to CdsDNA ) can be obtained and its plot versus the concentration of the
com-plementary DNA in solution, [cDNA] , is shown in Figure 3.6. Fitting such data, one can further compute the anity constant, KA = kkon
of f , for DNA-hybridization. If we
suppose in fact, that the DNA-pairing on the gold surface can be described by Langmuir adsorption isotherm [48, 72], we obtain the following equation that connects the per-centage change ofCd, the varying concentration of the complementary DNA in solution,
[cDNA], andKA [64, 67], dotted black curve:
Cd% = 1 + KKA · [cDN A]
A · [cDN A] (3.4)
An anity constant for DNA hybridization KA ' (0.35 ± 0.09) × 10 9 M−1,
cor-responding to an equilibrium dissociation constantKD = K A−1 ≈ (2.8 ± 0.7) nM, was
strands of similar length were adsorbed on a gold surface.
From data of Figure 3.6, a limit of detection (LOD) of 1 pM for the device is de-termined. This value is the smallest measured capacitance variation that can be un-ambiguously assigned to the hybridization process.Lower concentrations of [cDNA]in fact, lead to variations of Cd comparable to the control measurements,obtained for
non-complementary strands in solution at high (1 µM) concentration. In gure this region is highlighted with a solid line comprised between the two dotted red lines that represent three standard deviations around this value.We think that once we will have improved our experimental setup (especially the functionalization step by employing gold nanoparticles,for example) the detection limit of these devices for DNA-hybridization will be pushed towards the fM-range. Nevertheless,the LOD obtained with our device is already a very good starting point for further measurements in more complex human environments.
Obviously so far we made the assumption that the DNA-hybridization follows a Langmuir kinetics. In the next section we discuss the applicapibility and the limit of this approximation.
3.3 Kinetics and Dynamics of DNA Hybridization
In order to get insights into the DNA hybridization kinetics, we t our data with the well-known Langmuir adsorption model.In this model the adsorption of adsorbate (A) onto the surface of the adsorbant or solute (S) can be described using the following three assumptions, (see Figure3.7):
The surface of the adsorbant, assumed homogeneous, is in contact with a solution
containing an adsorbate which is strongly attracted from the surface;
The surface has a specic number of sites where the solute molecules can be ad-
sorbed and all sites are equivalent;
The adsorption involves the attachment of only one layer of molecules to the sur-
face, i. e. monolayer coverage only;
